Sure, that makes sense.  Let's take the most serious problems society faces and then see if we can find enough altruistic individuals that are willing to fix them for nothing.

I maintain that the solution to America's science problem is not to throw tax dollars at the issue (as the recent $24,000 per car 'cash for clunkers' program showed, governments do not accomplish even simple things well) but ease up on visa restrictions - we are letting students learn here and then not letting them work here as part of some weird protectionist scheme, and then wondering when other countries will overtake us.

There is an economic disincentive too but that is because there are too many people who want too few academic research jobs - the private sector is always hiring but somewhere in the 1970s there developed a myth that private companies don't do research unless it is sales focused.   

I'm not sure that visa restrictions will help much.  Take the situation with H1B visas now for technical fields.  It isn't that there is a shorter of qualified people, it's just that companies know that they can bring in people from overseas at half the price.  It short circuits the supply/demand relationship for U.S. workers and results in unemployment here while companies exploit a foreign labor pool.

In the end, it becomes a race to the bottom.  If more foreign visas were granted, there would be even less pay in the sciences.

Not to overlook the benefit society gains from finance and economics majors.

"How should we capture the smartest upcoming minds and recruit them for science?" is the wrong question, particularly in a free society. The results of science should speak for themselves, just as the results of any other discipline.

In Not Everyone Gets a Trophy, Bruce Tulgan presents years of qualitative research comparing up and coming generations of workers with older ones. One of his best points is that younger generations have highly advanced BS detectors. A wide-scale science careers PR campaign is as effective as painting "say no to drugs" on the sides of buses. It is weak, like asking people to buy cars to support America rather than to buy a good car.

Furthermore, the risks of the *success* of such a campaign may even be worse than the risks of failure, lest we find ourselves with too many scientists and too few financial analysts. At the least, if we are going to promote science, why not use science as a tool itself and scientifically demonstrate *why* more scientists are needed? Simply saying that they have lots of opportunities to help society is like saying that you like your hypothesis.